Proposition #1: Details matter

Proposition #2: People hear what they expect to hear

Lemma #1: People often misunderstand details

Theorem: Foreign aid agencies continue to use tranched operations even when a small modification would work better.

I leave the proof to the reader, but this theorem came to mind during recent discussions about Cash on Delivery Aid (COD Aid) as it would apply to financing primary education.

One important part of the COD Aid proposal (okay, it may seem like a detail, but bear with me) is that foreign funders would pay a fixed sum for each additional child who completes primary school. That means an incremental payment: if 10,000 more students complete school, you get paid X times 10,000. If 11,126 students complete school, you get paid X times 11,126. If only 9,456 get through, then it is X times 9,456.

Get the idea?

Explaining this to intelligent people who have worked for years in the aid community is a more difficult task than one might think because, for better and worse, the idea of paying for results has gotten tied up with the idea of tranches – payments that are made as a lump sum when a target is achieved. The use of tranches is among the most innovative changes to foreign aid since WWII, introducing the possibility of making payments that were no longer linked to specific project inputs. The most prominent of these operations were the structural adjustment loans of the World Bank in the 1980s; however, subsequent innovations have disbursed tranches against targets for maintaining social safety net spending through a crisis, implementing particular fiscal policies, and even making progress on social indicators.

Disbursing tranches against achieving some target is quite distinct from a traditional project loan, but it presents several drawbacks in practice. First, setting the target is quite difficult because of competing interests. Depending where you sit, you may be interested in a lower threshold to assure that the disbursement takes place or in a higher threshold to create pressures to improve performance. Note that the competing interests are not just between funding agencies and recipient countries because the funding agencies, themselves, include people and departments with both types of interest.

The second drawback is that if the target is not met, funders and recipients tend to re-enter negotiations over what to do – a kind of “Tranche Warfare.” Was the target unreasonable to begin with? Were the actions taken appropriate but the theory of change faulty? Should the condition be waived since other targets in the same operation were met? Any number of questions can be asked and debated, undermining the original intention. In fact, despite all the hue and cry over World Bank and IMF conditionality, few countries are ever sanctioned for failing to meet targets.

So what does this have to do with COD Aid?


Invariably, when people in the aid community hear that a COD Aid proposal will pay for the number of children who complete primary school, they think in terms of tranches. That would mean the funder and recipient negotiate a target (an extra 10,000 kids complete school), design a program to achieve progress, assess whether the target was met, and make payment if at least 10,000 kids get through.

But that is not the COD aid proposal. If we get out of tranche-thinking long enough to really think about what it means to pay a fixed amount for each child, the dynamics change substantially. It may seem like a detail, but it means:

  1. We don’t need to negotiate the target beforehand because the payment is per student (see illustration above). All that we need is the per-student ‘fee’ and an agreement on the units of progress.
  2. When the number of students is reported (and after verification, see COD Aid FAQs, there is no argument over what to do because there is no failure or success, there is only more success (e.g. 11,126 additional students) or less success (e.g. 9,456).

In other words, even without adopting a full COD Aid approach, many aid operations could change the dynamics of foreign aid substantially today by programming themselves to pay per unit of progress rather than for a binary pass/fail target. The European Union’s variable tranche budget support programs are an example that would be particularly easy to modify in this way.

So, is there anyone out there interested in getting out of the tranches?